Thanks for reaching out.
Your form submission was successful.
Close windowYour form submission was successful.
An error occurred while submitting the form.
Please try again.
Please try again.
All of us who work in information technology endure some level of repetitive work in our daily jobs. It can take the form of sending certain types of emails, routinely logging in to check the status of information on a website, or performing perfunctory tasks in a productivity application. Some information-processing jobs involve a great deal of repetition. You’ll see this across a variety of corporate departments, like with Accounts Payable/Receivable in Finance or new employee on-boarding in HR or new vendor setup in Procurement. You’ll also see these repetitive processes vary by industry, as with new loan processing in financial services or claims processing in health care.
The recent meteoric rise of RPA (Robotic Process Automation) technology has been fueled by a promise to automate many of these highly repetitive processes by leveraging the latest advancements in AI and machine learning. But evidence suggests many RPA solutions and deployments have fallen short of their promise. EY, for example, noted it has seen as many as 30-50 percent of initial RPA projects fail, citing process complexity and the lack of process understanding for many of the failures. A wise person once said, “you can’t fix what you can’t see.” Many RPA projects are trying to automate processes that aren't fully understood.
It is this very lack of understanding and visibility into the execution of an organization’s business processes that impedes not only automation initiatives but, more importantly, holds back the potential of the human workforce and ultimately the business’s bottom line. We knew there had to be a better way.
We founded Zeitworks to empower companies with the critical knowledge and insights they need to streamline their processes and make the most of their workforce. We believe that process discovery and our Dynamic Process Improvement™ technologies are the foundational step towards understanding and optimizing the execution of your business processes. Whether you’re new to process improvement or a grizzled veteran of process automation, Zeitworks technology helps to shine a light on the hard-to-see inner workings of your business. The result is analytics and actionable insights that empower you to make data-driven decisions. In the end, with Zeitworks, your team has the information you need to remove bottlenecks and inefficiencies, establish standardized benchmarks, comply with regulations, drive predictable automation initiatives, and continuously learn and improve your bottom line. Want to learn more? Then read on!
I have been a serial technology entrepreneur in Seattle for more than 20 years, having focused my career on building innovative data and machine learning-driven products and services across several industries and verticals. About five years ago I joined Madrona Venture Labs, a startup studio in Seattle, where I was CTO and Managing Director. We worked on a lot of B2B, data-driven machine learning ideas, products, and companies. One of those ideas and companies was Zeitworks.
Excited by recent investments in the rapidly growing RPA space, we started exploring the RPA ecosystem in early 2019, interviewing RPA vendors, customers, management consulting firms, and system integrators. It was during that research that we realized there is a huge problem and opportunity upstream from automation and RPA.
You see, before a repetitive business process can be automated, improved, or re-engineered, it must first be ‘discovered’, quantified, and understood. It turns out that gaining insight into one’s existing business processes is not so easily accomplished. That’s where Zeitworks comes in.
We learned that many teams of information processing workers performing repetitive tasks (from loan processing and accounts payable/receivable to new customer onboarding) lack knowledge and insight into the nature of those processes. The official way to perform them is often not documented, or at best out of date, and there is a lack of insight into the levels of efficiency. It is not uncommon for employees to execute their work uniquely, often not sharing best practices or tribal knowledge. This inefficiency costs companies tremendous amounts of time and money and leaves managers and process designers with their hands tied, lacking data to drive improvements, efficiencies, and even automation initiatives.
In our research we heard CIOs and managers tell us they wanted to know, “where am I right now?”, and “how do I gain transparency into the efficiency level of the resources I have?” We also heard that yes, automation is interesting, but there are more basic issues and questions:
Of course, we immediately wondered, how has ‘process discovery’ been done historically? The answer was apparent - in a word, manually. Management consulting agencies and system integrators are hired to unravel the riddle of how a process is constructed. This typically entails bringing in groups of junior associates who stand over the shoulders of employees to scribe steps they see them take on their desktop computers as they go about their work. After weeks of observation and manual collation of information, a ‘process map’ (that only represents a single point in time), is produced and referenced as a benchmark for analysis, improvements, documentation, and training.
In more recent years, as RPA began to rise, the pressure to find more automated and intelligent technology solutions to replace these manual, slow and expensive process illumination projects has become of paramount importance and value.
It turns out that ‘process discovery’ (or task mining, as it is sometimes called by RPA vendors) is a really hard problem to solve with technology in an automated and intelligent fashion. Even highly repetitive or cyclical business processes can be complex and messy and contain significant variances. Humans are imperfect. They tend to drift from structured execution organically over time. Sometimes business processes are deliberately designed and planned, other times they are started by a single person who feels their way through a set of requirements or obvious steps. They then go on to train new people, who add their own touch, spin, or preferences to the execution. Before long, everybody is doing their own thing, and nobody really understands the right and the best way to execute the processes. This lack of transparency and inherent inefficiency costs companies critical time and money. It also impedes automation efforts.
Adding to this challenge is that many repetitive information processing tasks are done primarily by individuals, working across multiple applications on their desktop computers. Older (legacy) process mining tools provide high-level insights into cross-organizational processes that run on enterprise systems (e.g. ERP, CRM) by mining structured log files.
But many processes that are undocumented and inefficient require a different approach and solution to the problem. To measure individuals working on multi-application processes, signals need to be gleaned from a user’s interactions with application user interfaces and not just neatly distilled log files. The data Zeitworks collects is therefore a verbose continuous stream of low-level user events and gestures collected passively and unobtrusively as employees go about their work.
Collecting the data is just one part of the challenge of process discovery. Making sense of this data is the paramount challenge.
All of us at Zeitworks are veteran technologists and entrepreneurs. Many of us have been working with machine learning technologies and the myriad of data challenges that accompany them in real-world business and consumer applications for many years across many companies. We’re living in a golden age for machine learning and its business and consumer applications, which is changing the way we live and work in both worlds. For our technologists and data scientists at Zeitworks, it’s a great time to be trying and leveraging both classic sequential models (e.g. HMMs) as well as the newest deep learning technologies.
But machine learning models, architectures, and algorithms will come and go. We know from experience that data is the new oil for companies- especially those going through Digital Transformation. A deep understanding of the data, along with clever data preparation, is key to extracting value and applying any kind of predictive model successfully. We see Zeitworks fundamentally as a data company. Our sensors passively measure your most valuable assets, the integrity of the execution of your business process, and your workforce. The highly valuable rich data we capture and accumulate is the fuel you need to make you and your company better and smarter.
After incubating Zeitworks inside Madrona Labs, I continued to advise the company for a year before I joined Zeitworks as CEO in June of 2021. I am excited to be operating again after five years of working in a lab. Early-stage technology companies are where I thrive. I love scrappy and driven teams with everyone owning big pieces of the technology, product, and operations. I love the opportunity to work closely with customers, moving quickly and aggressively, to build an important, and valuable product and service we hope can be incredibly helpful and transformative to you and your organization. We’d love to hear from you [link to contact us], give you a tour of our technology and product, and help you get started getting the visibility and insights you need to start transforming the efficiency of your business and human workforce today!
